A/B Testing Core Web Vital Calculator Examples
See how to interpret control-versus-variant Core Web Vitals results for common A/B testing performance scenarios.
These worked scenarios illustrate how individual LCP, INP, and CLS changes can affect the combined Core Web Vitals index. Compare like-for-like data sources and segments so the result reflects the variant rather than a change in measurement conditions.
Balanced improvement across all three metrics
The control records 2.8 s LCP, 240 ms INP, and 0.12 CLS. The variant records 2.4 s LCP, 180 ms INP, and 0.08 CLS.
Input Summary
Control metrics
2.8 s LCP, 240 ms INP, 0.12 CLS
Variant metrics
2.4 s LCP, 180 ms INP, 0.08 CLS
Calculation Breakdown
- 1Individual changesLCP: 14.3%; INP: 25.0%; CLS: 33.3%All positive
- 2Combined index comparisonControl: 1.17; Variant: 0.8924.4% improvement
Result Summary
Combined index comparison
24.4% improvement
A/B Testing Core Web Vital Calculator
The variant improves every included metric and reduces the combined index by about 24.4%.
Faster loading but worse responsiveness
The control records 3.0 s LCP, 180 ms INP, and 0.09 CLS. The variant records 2.5 s LCP, 240 ms INP, and 0.09 CLS.
Input Summary
Control metrics
3.0 s LCP, 180 ms INP, 0.09 CLS
Variant metrics
2.5 s LCP, 240 ms INP, 0.09 CLS
Calculation Breakdown
- 1Individual changesLCP: 16.7%; INP: -33.3%; CLS: 0.0%Mixed result
- 2Combined index comparisonControl: 1.17; Variant: 1.078.6% improvement
Result Summary
Individual changes
Mixed result
A/B Testing Core Web Vital Calculator
The combined score improves modestly, but the variant has a meaningful responsiveness regression.
Better stability on a content-heavy page
The control records 2.4 s LCP, 190 ms INP, and 0.18 CLS. The variant records 2.4 s LCP, 190 ms INP, and 0.06 CLS.
Input Summary
Control metrics
2.4 s LCP, 190 ms INP, 0.18 CLS
Variant metrics
2.4 s LCP, 190 ms INP, 0.06 CLS
Calculation Breakdown
- 1Individual changesLCP: 0.0%; INP: 0.0%; CLS: 66.7%CLS-led improvement
- 2Combined index comparisonControl: 1.24; Variant: 0.8432.3% improvement
Result Summary
Combined index comparison
32.3% improvement
A/B Testing Core Web Vital Calculator
The variant's improvement is driven entirely by reduced layout shift.
How to Read Your Results
A positive LCP, INP, or CLS improvement percentage means the variant's value is lower than the control's.
A negative percentage indicates a regression because the variant metric is higher.
Compare the three individual results before relying on the overall index.
An index below 1.00 means the average normalized metric is below the calculator's selected reference thresholds.
Treat close differences carefully because ordinary measurement variation can affect Core Web Vitals data.
Assumptions & Important Notes
- All examples use the same measurement type and comparable user segments for the control and variant.
- The combined index uses equal weighting for normalized LCP, INP, and CLS.
- Reference thresholds are 2.5 seconds for LCP, 200 ms for INP, and 0.1 for CLS.
Related Examples
Frequently Asked Questions
What is a good example of a positive Core Web Vitals A/B test result?
A positive result has lower variant values for the metrics that matter, with a lower combined index and no important guardrail regression.
Can an A/B test have a better LCP but a worse INP?
Yes. Changes that reduce loading work can still add JavaScript, event handling, or other work that affects interaction responsiveness.
Should I roll out a variant when the overall index improves?
The index is one directional performance signal. Consider data quality, uncertainty, individual metric changes, and the experiment's separate success criteria.
Why can CLS make a large difference to the index?
CLS is normalized against 0.1, so a change from 0.18 to 0.06 is a large relative reduction against that reference value.
Ready to calculate your own result?
Use the live calculator with your own inputs, timing, and preferences.